Why Your Tracked Rankings Don't Match What You See in Google

Your rank tracker says your page is in position 4. You search the keyword yourself and find it at 9, or on page two, or in a spot your tracker never mentioned. Both can be correct.

A Google ranking depends on who searched, where, on what device, and when. A tracker controls those conditions and holds them steady. Your own search uses whatever your browser, account, and network happen to send. In 2026 there is also a newer cause: Google sometimes gives automated requests a different page than it gives people.

These are the usual reasons the numbers disagree, roughly from most to least common.

1. Location

Location is the biggest cause. Google localizes far more queries than most people expect, including queries with no obvious local intent. A search for "how to file for divorce" returns a different state's court websites depending on where you search.

Since April 2025, the country domain no longer changes anything either. Google redirects country domains to google.com and localizes by location instead. A tracker set to "United States" may be sampling one city, or a broad national location that matches no real person. You are searching from wherever you are.

Our localization guide explains the gl, hl, and uule controls that decide which version of a SERP a request receives. To compare fairly, set the tracker to your city, or search from the tracker's location.

2. Device

Mobile and desktop results differ. They often have different SERP features and different organic order. Most trackers check one device per keyword. If you check on a laptop and the tracker checks mobile, expect differences.

3. Your account and browser

When you are signed in, Google can tailor results to your activity. Even signed out, your browser sends a language preference, keeps cookies from earlier searches, and may have SafeSearch settings. Sites you visit often can rank higher for you than for a stranger.

A private window helps, but it doesn't make you a neutral searcher. Your location and network still apply.

4. How position is counted

"Position 4" can mean two things:

• Organic rank. Your place among the classic organic results, ignoring ads and SERP features.
• Absolute position. Your place among everything on the page, including AI Overviews, People also ask, local packs, and video carousels.

A page can be fourth among organic results and still sit below the fold under an AI Overview, a local pack, and four ads. Check which number your tracker reports. In the PrismCrawl response, each result has both a rank and a position, so you can report whichever matches your question.

Positions past page one add another wrinkle. Each page of results is a separate request, and Google doesn't always return exactly 10 organic results per page. If a tracker numbers results as the page offset plus the position on the page, a page with 9 results makes the next page look like it skipped a spot, and a page with 11 makes two URLs share a position. Pages fetched separately also overlap: a URL at #10 often shows up again as #1 on the next page. A tracker should merge pages in order, drop URLs it has already seen, and renumber the combined list from 1.

Search Console uses its own rules. Its average position blends every query, device, and location where your page appeared, and Search Engine Roundtable reports that all links in an AI Overview share one position. It will rarely match a single tracked check.

5. Timing

Rankings move during the day, and much more during Google updates. Google's Search Status Dashboard listed core updates in March and May 2026 and spam updates in March, June, August, and September 2026. A tracker that checked at 6:00 UTC and a person who searched at 18:00 UTC may have seen different pages. Our guide to telling a Google update from a data problem covers how to read movement during an update.

6. Which network the check came from

Google doesn't always serve the same ranking to every network. In our own testing, the same query, location, and device sometimes returned a noticeably different organic order depending on where in a proxy network the request went out, often one led by directories. Changing cookies, browser version, or location settings didn't remove it.

The variance is strongest on location-specific searches, such as service businesses ("plumber in [city]", "roof repair near me"). National and informational keywords vary too, but less.

A tracker that spreads checks across a large proxy network can see a site flip between its usual position and missing from one check to the next. A tracker that always uses the same few connections looks steadier, but only because it keeps seeing one version of the ranking.

For organic-only tracking, Google's Web filter (udm=14) largely removes the difference. In a test of 10 keywords across pages 1 to 5, top-10 agreement between repeated checks rose from about half to over 80%, every page returned exactly 10 organic results, and a site that usually ranked in the top 10 went missing from the top 50 in 3% of checks instead of 11%. The trade-off is that the Web filter drops SERP features such as AI Overviews, local packs, and People also ask, so track those with a separate, unfiltered check.

7. Bad data in the tracker

Since September 2026, the tracker can also be wrong in ways that look right. Google answers some automated requests with poisoned SERPs, where most of a keyword's usual sites are swapped out, or with partial query results that answer only one word of the query. Both return pages that look normal, and both usually disappear on the next check.

Earlier problems still show up too. A tracker that never handled Google's /goto links may record google.com as the ranking domain, which makes your page look like it vanished.

How to compare a tracked ranking with your own search

When the numbers disagree and you need to know which one to trust:

1. Match the location. Use the same city or postal code the tracker uses.
2. Match the device. Check on mobile if the tracker checks mobile.
3. Search signed out in a private window, with the same language setting.
4. Compare like with like. Count organic results only, or count everything, the same way your tracker does.
5. Check the time. Look at the tracker's check time and whether an update was running.
6. Look at the whole page the tracker saw. If your tool stores the raw HTML or the full result list, compare the domains with what you see. If most of them differ, the tracker probably received a poisoned page.
7. Check again before you report. A second check a few minutes later settles most oddities that appear only once.
8. For organic positions, check with udm=14. It keeps repeated checks consistent across networks, at the cost of SERP features.
9. Use Search Console as the tiebreaker for trends over days or weeks. It reflects what real searchers saw.

If you build your own tracking on a SERP API, store the location, device, check time, and request ID with every observation. Our rank tracker build guide shows a schema that keeps those fields together.

Frequently asked questions

Why is my Google ranking different from my rank tracker?

The most common reasons are location, device, and personalization. Google localizes many queries, mobile and desktop results differ, and searches made while signed in can be tailored to you. Trackers also count position differently, and since September 2026 some automated checks receive poisoned results.

Why do two identical rank checks return different positions?

Google doesn't always serve the same ranking to every network, especially for location-specific searches, and each check, or each page of a check, may go out through a different part of a proxy network. For organic-only tracking, Google's Web filter (udm=14) keeps repeated checks much more consistent. When you collect several pages, merge them, drop repeated URLs, and renumber the list.

Which is more accurate, my own search or a rank tracker?

Neither is automatically right. Your own search shows what you see. A tracker that is set up carefully shows what a searcher at a set location and device sees, consistently over time. For trends, compare the tracker with Google Search Console, which reflects real searchers.

Does incognito mode show unbiased Google rankings?

It removes some personalization, but your location, network, and language settings still affect results. To match a tracker, set the same location and device as well.

Why does Search Console show a different average position?

Search Console averages your position across every query, location, and device where your page appeared, and links in an AI Overview share one position. A single tracked check measures one query at one location at one time, so the two rarely match exactly.